Electronic Thesis and Dissertation Repository


Doctor of Philosophy


Chemical and Biochemical Engineering


Dr Ajay Kumar Ray


This thesis is concerned with application of mechanistic models for recovery and purification of two minor milk proteins to develop an efficient and robust process. A fundamental and quantitative understanding of the underlying mechanisms assists to evaluate chances and challenges in non-linear chromatography.

The first chapter considers adsorption isotherm data of two minor whey proteins on cation exchanger under various conditions and used as the basis to develop a predictive approach for correlating adsorption behavior using a mechanistic isotherm model. The SMA isotherm model explicitly considers the contributions of protein-adsorbent and protein-protein interactions in the simulation of salt gradients in ion exchange chromatography.Sensitivity and robustness analysis by factorial design of experiments within this framework showed to be highly consistent and even allowed for upscale predictions with an excellent quality.

In the next part of the thesis, the nonlinear gradient elution was to be optimized by three process factors the length of gradient, final salt concentration at the end of gradient and flow velocity. Predictions based on response surface modeling (RSM) approach were applied to reveal significant process factors. The optimal operating point was then determined by calibrated mechanistic model within and outside the design space. The operating conditions containing optimal information were experimentally verified which confirmed simulations accuracy.

The third chapter considers the effects of scale-up and operating conditions on dynamic adsorption of proteins. For two columns having similar bed height, flow distribution properties was observed under non-binding conditions. Elution profiles were employed to determine dominant mass transport mechanisms. Breakthrough profiles were compared at different flow rates and protein loading concentrations.The efficiency of the columns in terms of HETP and dynamic binding capacity were calculated and compared for two columns.

The outcomes resulting from the application of mechanistic models to the purification of lactoperoxidase and lactoferrin in this thesis exploit the platform for the next step towards the recovery of high-value proteins at industrial scales.